ARE WE CLOSING THE SCHOOL DISCIPLINE GAP?



Similar documents
THE FUTURE OF HIGHER EDUCATION IN TEXAS

The Disparate Impact of Disciplinary Exclusion From School in America

ANTHONY P. CARNEVALE NICOLE SMITH JEFF STROHL

Foreign Language Enrollments in K 12 Public Schools: Are Students Prepared for a Global Society?

LexisNexis Law Firm Billable Hours Survey Top Line Report. June 11, 2012

2016 Individual Exchange Premiums updated November 4, 2015

Mapping State Proficiency Standards Onto the NAEP Scales:

50-State Analysis. School Attendance Age Limits. 700 Broadway, Suite 810 Denver, CO Fax:

STATE INCOME TAX WITHHOLDING INFORMATION DOCUMENT

Closing the College Attainment Gap between the U.S. and Most Educated Countries, and the Contributions to be made by the States

Hail-related claims under comprehensive coverage

Incarcerated Women and Girls

Rates are valid through March 31, 2014.

VCF Program Statistics (Represents activity through the end of the day on June 30, 2015)

How To Regulate Rate Regulation

Health Insurance Price Index Report for Open Enrollment and Q May 2014

The Nation s Report Card

INTRODUCTION. Figure 1. Contributions by Source and Year: (Billions of dollars)

Health of Wisconsin. Children and young adults (ages 1-24) B D. Report Card July 2010

Alabama Commission of Higher Education P. O. Box Montgomery, AL. Alabama

Health Insurance Exchanges and the Medicaid Expansion After the Supreme Court Decision: State Actions and Key Implementation Issues

U.S. Department of Housing and Urban Development: Weekly Progress Report on Recovery Act Spending

United States Bankruptcy Court District of Arizona NOTICE TO: DEBTOR ATTORNEYS, BANKRUPTCY PETITION PREPARERS AND DEBTORS

APPENDIX B. STATE AGENCY ADDRESSES FOR INTERSTATE UIB CLAIMS

Notices of Cancellation / Nonrenewal and / or Other Related Forms

United States Bankruptcy Court District of Arizona

Table 1: Advertising, Marketing and Promotional Expense as a Percentage of Net Operating Revenue

How To Compare Ehealth To A Health Insurance Plan

NYCOM 2009 Entering Class - Matriculant Comparison Data

Health Coverage for the Hispanic Population Today and Under the Affordable Care Act

NAAUSA Security Survey

kaiser medicaid and the uninsured commission on The Cost and Coverage Implications of the ACA Medicaid Expansion: National and State-by-State Analysis

STATE SUPPLEMENTAL NUTRITION ASSISTANCE PROGRAM PARTICIPATION RATES IN 2009 FOOD AND NUTRITION SERVICE

The Youth Vote in 2012 CIRCLE Staff May 10, 2013

Alaska (AK) Arizona (AZ) Arkansas (AR) California-RN (CA-RN) Colorado (CO)

Health Insurance Coverage of Children Under Age 19: 2008 and 2009

Annual Survey of Public Employment & Payroll Summary Report: 2013

ENS Governmental Format Status (As of 06/16/2008)

The Case for Change The Case for Whopping Big Change

COMMERCIAL FINANCE ASSOCIATION. Annual Asset-Based Lending and Factoring Surveys, 2008

Regional Electricity Forecasting

The Survey of Undergraduate and Graduate Programs in Communication. Sciences and Disorders has been conducted since Surveys were conducted in

How To Get A National Rac (And Mac)

How To Rate Plan On A Credit Card With A Credit Union

LIMITED PARTNERSHIP FORMATION

Piloting a searchable database of dropout prevention programs in nine low-income urban school districts in the Northeast and Islands Region

Health Insurance Coverage: Early Release of Estimates From the National Health Interview Survey, January June 2013

Resource Brief: Ombudsman Program Data Management Systems

FR Y-14Q: Retail US Auto Loan Schedule Instructions

Estimating college enrollment rates for Virginia public high school graduates

ASNT Conference Edition - Fall Salary Survey Your Path to the Perfect Job Starts Here.

AAIS Mobile-Homeowners 2008 Series

U.S. Department of Education NCES NAEP. Tools on the Web

Health Insurance Coverage: Early Release of Estimates From the National Health Interview Survey, 2013

Community College/Technical Institute Mission Convergence Study

ehealth Price Index Trends and Costs in the Short-Term Health Insurance Market, 2013 and 2014

State Insurance Information

High School Graduation in Texas

TAX PREP FEE PHILOSOPHY. Copyright 2013 Drake Software

Cancellation/Nonrenewal Surplus Lines Exemptions

Benefits of Selling WorkLife 65

Cancellation of Debt (COD) R. Bruce McCommons Harford County, MD TrC 12/4/2013

CPT Codes For Spirometry

COUNCIL OF GRADUATE PROGRAMS IN COMMUNICATION SCIENCES AND DISORDERS NATIONAL SURVEY OF UNDERGRADUATE AND GRADUATE PROGRAMS

A descriptive analysis of state-supported formative assessment initiatives in New York and Vermont

Enrollment Snapshot of Radiography, Radiation Therapy and Nuclear Medicine Technology Programs 2013

States Future Economic Standing

STATE HOMELESSNESS. The. An examination of homelessness, economic, housing, and demographic trends at the national and state levels.

Cost and Benefits of Individual and Family Health Insurance. December 2013

National Bureau for Academic Accreditation And Education Quality Assurance PUBLIC HEALTH

NHIS State Health insurance data

CIVIL RIGHTS DATA COLLECTION

Required Minimum Distribution Election Form for IRA s, 403(b)/TSA and other Qualified Plans

State Technology Report 2008

LICENSED SOCIAL WORKERS IN THE UNITED STATES, 2004 SUPPLEMENT. Chapter 2 of 5. Who Are Licensed Social Workers?

ADEA Survey of Dental School Seniors, 2014 Graduating Class Tables Report

How To Know The Nursing Workforce

Copyright 2005 by The Schott Foundation for Public Education

New York Public School Spending In Perspec7ve

How NYPIRG s Physician Supply Report Misses the Mark. September 2014

ONLINE SERVICES FOR KEY LOW-INCOME BENEFIT PROGRAMS What States Provide Online with Respect to SNAP, TANF, Child Care Assistance, Medicaid, and CHIP

Alaska (AK) Arizona (AZ) Arkansas (AR) California-RN (CA-RN) Colorado (CO)

Department of Banking and Finance

Radiologic Sciences Staffing and Workplace Survey 2015

CDFI FUND NEW MARKETS TAX CREDIT PROGRAM:

Health Coverage by Race and Ethnicity: The Potential Impact of the Affordable Care Act


2: & # CMO (% # EMO

REPORT: DEMOGRAPHIC SURVEY OF UNDERGRADUATE AND GRADUATE PROGRAMS IN COMMUNICATION SCIENCES AND DISORDERS

Florida Workers Comp Market

Motor Vehicle Financial Responsibility Forms

Issuers Owing Refunds for 2013

NEW CARRIER SIGN UP REQUEST FORM

Home Schooling Achievement

Enrollment Snapshot of Radiography, Radiation Therapy and Nuclear Medicine Technology Programs 2012

NCCI Filing Memorandum Item B-1420

AAIS Personal and Premises Liability Program

Economic and Personal Finance Education in Our Nation s Schools

PEER Analysis of OSHA Recordkeeping Inspections Done Pursuant to its National Emphasis Program (NEP)(10/09-8/10) SUMMARY OF DATA

*Time is listed as approximate as an offender may be charged with other crimes which may add on to the sentence.

Transcription:

ARE WE CLOSING THE SCHOOL DISCIPLINE GAP? THE CENTER FOR CIVIL RIGHTS REMEDIES F E B R U A R Y, 2 0 1 5 By Daniel Losen, Cheri Hodson, Michael A. Keith II, Katrina Morrison, and Shakti Belway

Acknowledgements A special thank you to Atlantic Philanthropies for funding this research and to The California Endowment for their generous assistance with the final edits and finishing process. The authors of this report are especially grateful to Tia Elena Martinez whose guidance organizing the coding of the data and templates for analysis made this work possible, as well as Valerie Okelola who made important contributions and provided a good deal of the initial data crunching and data comparisons work. We are also extremely fortunate to have had the technical editing assistance of Dody Riggs, who provided extensive edits to each component of the report and district profiles. A warm thank you to Laurie Russman, the administrative point person at the Civil Rights Project for the numerous ways her work supported these efforts. We would also like to thank Norman Black of Hatcher Communications for all of his help highlighting the many newsworthy aspects of this report. Our acknowledgments would not be complete without a deep thank you to the Directors of the Civil Rights Project, Gary Orfield and Patricia Gándara, whose particular feedback inspired several improvements and whose enduring support and wisdom is a constant source of inspiration for the work of the Center for Civil Rights Remedies. 2

Table of Contents: 1 I. Introduction and National Overview II. Methods III. Elementary Level Findings a. National Scope of Suspension Use b. State Rankings c. District Highlights IV. Secondary Level Findings a. National Scope of Suspension Use b. State Rankings c. District Highlights V. Conclusions and Recommendations a. For Media b. For Policymakers c. For Educators d. For Parents and Advocates VI. References and Endnotes VII. Appendix A: K-8 schools VIII. Appendix B: Details on Methods and Data Cleaning and preferred measures of racial disparities. IX. Appendix C: Details on most improved large districts. 1 Profiles for 24 district profiles available at http://tinyurl.com/ccrrnationalreports. Profiled districts are: Atlanta (GA), Columbus (OH), Dallas (TX), Houston (TX), Fairfax (VA), Memphis (TN), Pittsburg (PA), Boston (MA), Albany (NY), Providence (RI), Dade County (FL), St. Louis (MO), Detroit (MI), Jefferson County (KY), Las Vegas [Clark County] (NV), Madison (WI), St. Paul (MN), New Haven (CT), Oklahoma City (OK), Baltimore City (MD), Montgomery (AL), Kansas City (KS), Washington (D.C.), and Seattle (WA). 3

ARE WE CLOSING THE SCHOOL DISCIPLINE GAP? Introduction and National Overview Nearly 3.5 million public school students were suspended out of school at least once in 2011-12. 12 That is more than one student suspended for every public school teacher in America. 3 This means that more students were suspended in grades K-12 than were enrolled as high school seniors. 4 To put this in perspective, the number of students suspended in just one school year could fill all of the stadium seats for nearly all the Super Bowls ever played!(the first 45). 5 Moreover, recent estimates are that one in three students will be suspended at some point between kindergarten and 12 th grade (Shollenberger, 2015). If we ignore the discipline gap, we will be unable to close the achievement gap. Of the 3.5 million students who were suspended in 2011-12, 1.55 million were suspended at least twice. Given that the average suspension is conservatively put at 3.5 days, we estimate that U.S. public school children lost nearly 18 million days of instruction in just one school year because of exclusionary discipline. 67 Loss of classroom instruction time damages student performance. For example, one recent study (Attendance Works, 2014) found that missing three days of school in the month before taking the National Assessment of Educational Progress translated into fourth graders scoring a full grade level lower in reading on this test. New research shows that higher suspension rates are closely correlated with higher dropout and delinquency rates, and that they have tremendous economic costs for the suspended students (Marchbanks, 2015), as well as for society as a whole (Losen, 2015). Therefore, the large racial/ethnic disparities in suspensions that we document in this report likely will have an adverse and disparate impact on the academic achievement and life outcomes of millions of historically disadvantaged children. This supports our assertion that we will close the racial achievement gap only when we also address the school discipline gap. Figure 1. Elementary and Secondary Out-of-school Suspension Rates by Subgroup, 2011-12 Elementary Secondary 25% 23.2% 20% 15% 10% 5% 0 2.6% 10.1% All 2.9% 11.9% American Indian 1.2% 7.3% Hawaiian/ Pac. Isl. 2.5% 0.5% 7.6% 10.8% 2.1% 1.6% 6.7% 1.5% 11% Asian Black Latino White English Learner 5.4% 18.1% With Disability 4

Suspension rates typically are three to five times higher at the secondary level than at the elementary level, as illustrated in figure 1. Furthermore, the actual size of the racial gap, such as that between Blacks and Whites, is much greater at the secondary level. The national summary of suspension rate trends for grades K-12 indicates that these rates increased sharply from the early 1970s to the early 2000s and then more gradually, until they leveled off in the most recent three-year period (see figure 2). We conclude that in this recent period, no real progress was made in reducing suspension rates for grades K-12. After many years of widening, the gap in suspension rates between Blacks and Whites and between Latinos and Whites narrowed slightly in the most recent time period that is, the 2009-10 and 2011-12 school years. The gap narrowed, however, only because of the increase in the White suspension rate. Specifically, 16% of Blacks and 7% of Latinos were suspended in both years, while rates for Whites rose from 4% to 5%. 8 Figure 2. Suspension Rates over Time by Race/Ethnicity, K-12 9 18% 16% 14% 12% 10% 8% 6% 4% 2% 0 6% 3% 3% 10% 5% 4% 13% 6% 5% 15% 15% 7% Black Latino 1972-73 1988-89 2000-01 2003-04 2005-6 2009-10 2011-12 7% 5% 5% White 16% 16% 7% 4% 7% 5% Data Source: U.S. Department of Education, Office for Civil Rights We next broke down the national trend analysis to the elementary and secondary levels. We only had the necessary data for the three years shown in figure 3. Despite the persistence of deeply disturbing disparities, the good news is that we estimate a slight reduction nationally in suspension rates for Blacks, Latinos, and Whites at the secondary level, along with a small narrowing of the racial discipline gap. However, this is countered by a small increase (less than one percentage point) in the use of suspensions at the elementary level and a slight increase in the racial gap at the elementary level. Breakdown by Elementary and Secondary Levels and by Additional Subgroups Provides a Clearer Picture This report breaks down the state and district data at the elementary and secondary levels to present a much clearer picture of current suspension practices and how they vary across similar grade configurations. 10 When we look at subgroups of children, we find that the most disturbing disparities reflect the disparate impact of suspensions on children who fall into more than one category. For example, 5

Figure 3. Suspension Rates over Time by Race/Ethnicity: Elementary and Secondary Levels 30% 24.3% 25% 20% 15% 12% 11.8% 10% 5% 0.4% 0 1972-1973 Black Secondary Latino Secondary White Secondary 1.2% 1.1% 0.7% 0.5% White Elementary 6.7% 6.6% 1.1% 1.4% Latino Elementary 10.8% 7.1% 6.1% Black Elementary 23.2% 2009-2010 2011-2012 when we look at racial and gender disparities at any grade level, the highest suspension rates typically are for Black males, followed by Black females and/or Latino males. However, the highest rates are found among secondary students with disabilities, which we further disaggregate by race and gender (see figure 4). Specifically, Black males are at the highest risk for suspension (33.8%), followed by Latino males (23.2%). Surprisingly, Black females with disabilities are suspended at higher rates (22.5%) than White males with disabilities at both the elementary and secondary level (figure 4). Figure 4. Suspension Rates for Students with Disabilities, Further Disaggregated by Race/Ethnicity and Gender Latino 1.9% White 1.1% Black 12.1% 6.6% 4.8% 7.3% 5.6% All 2.2% 0 5% 6.8% Students with Disabilities 16.2% 11.7% 10% 23.2% 22.5% 13.7% 15% Female Secondary Male Secondary 33.8% Female Elementary Male Elementary 21.4% 20% 25% 30% 35% 40% In this report, and in our earlier report based on the 2009-10 school year, we found that schools suspend students with disabilities at rates that are typically two to three times higher than for their non-disabled peers. By law, schools may not suspend students with disabilities for behavior that is caused by their disability. At first glance, the significant difference in suspension rates raises questions as to whether schools are failing to meet their legal and moral obligations to provide a free and appropriate public education to students with disabilities, in particular those who are frequently suspended. Moreover, the data on the 6

tremendous racial and gender disparities in the suspension rates of students with disabilities suggest that the rights of students with disabilities along the lines of race and gender are being unlawfully violated. State Rankings at the Elementary and Secondary Level Raise Serious Questions For the first time ever, we provide a ranking of suspension rates for 48 states. The data raise disturbing questions: Why did school authorities in Florida suspend 19% of all secondary students? Why does North Carolina suspend American Indian secondary students at such high rates (21%)? Why does Missouri suspend over 14% of Black elementary students and have the largest Black/White gap at the elementary level? Are Rhode Island s policies or education resources so deficient that they lead to more frequent suspensions? Why are so many (19%) secondary-level English learners suspended in Montana? It is worth noting here that English learners are particularly vulnerable in certain states. For example, when rank ordered for English learners three states with English-only instruction policies, Arizona, Massachusetts and California, were also among the highest 20 states for English learner suspension rates. District-Level Data Are Critical to Our Understanding It is critically important to note that suspension rates, and the size of the racial gap, vary dramatically from one district to the next. We believe that parents, educators, policymakers, and advocates need to know if children in their district are subject to excessive discipline. To meet this need, this report, together with the companion spreadsheet, provides analysis at both the elementary and secondary level for every district in the nation for school year 2011-12. We also provide a detailed trend analysis for every district that provided reliable data in 2009-10, which includes districts from Boston to St. Louis to Los Angeles. 11 Furthermore, we know from the latest studies that much of what determines whether a school is high or lower-suspending is directly influenced by its leaders (Skiba, 2015). In fact, most large districts show a great deal of variation in suspension rates from one school to the next. For example, a statewide study that tracked every middle school student in Texas, after controlling for race, poverty, and district policy, revealed that school-level factors had a tremendous impact on the suspension rate (Fabelo, 2011). Similarly, a study by Skiba et al. (2015) on the use of suspension throughout Indiana found that, after controlling for race and poverty and other significant factors, one variable stood out as the strongest predictor of both suspension rates and disparities in suspension by race: principals attitudes toward the use of harsh discipline. We also calculated the number of high- and lower-suspending schools for the nation and for each state. At the secondary level, any school or district that suspended 25% or more of any major racial/ethnic group s secondary enrollment was labeled high-suspending, and any school that suspended 10% or less of every major racial/ethnic group s secondary enrollment was deemed lower-suspending. When we aggregate our counts of high- and low-suspending schools up to the national level, we find, for example, that of the 34,000 secondary schools that met the criteria, 24% suspended one-quarter or more of at least one major racial or ethnic group. Although that figure may be alarming, we did find that more secondary schools (38%) were lower-suspending than high-suspending. At the elementary level, we changed the criteria. Any school or district that suspended 10% or more of any major racial/ethnic group s elementary enrollment was labeled high-suspending, and any school or district that suspended less than 2% of every major racial/ethnic group s elementary enrollment was deemed lower-suspending. We found that 37% of all elementary schools were lower-suspending and just 17% were high-suspending. In other words lower-suspending elementary schools outnumbered 7

high-suspending schools by more than two to one (approximately 16,765 to 7,520). Once again, the wide variation in elementary school suspension rates suggests that school policies and practices make a significant difference. However, according to both district-level research and the testimony of school superintendents and administrators in Baltimore, Los Angeles, and elsewhere (Losen, 2015), we also know that districts can foster effective school leadership and positive change at the school level. Notable district-level factors include whether or not schools discipline disparities are remedied, conducting careful selection and training of principals, providing support for teacher and leadership training, initiating changes to the school discipline code of conduct, and providing the specific behavioral supports and services that students with disabilities need. Furthermore, district policies can determine the degree to which the more effective practices are identified, promoted, and replicated (Gonzalez, 2015). Memphis, Tennessee: Huge Disparities, but Also Signs of Progress For each district in the nation, we report on discipline disparities by gender, disability status, and English learner status, and we further disaggregate the data by a cross-section of all these groups for 2011-12. Our district profiles offer good examples of what one can find out about a given district. Here we provide just a segment of one such profile, that of Memphis, Tennessee; the full profile is available in a companion document. We picked Memphis because it is among the highest-suspending large school districts in the nation, and because it reported reliable data for both 2009-10 and 2011-12, which allows us to describe the trends in its suspension rates. Figure 5 compares suspension rates in Memphis by race and disability status at both the elementary and secondary levels. Figure 5: Risk for Suspension at the Elementary and Secondary Levels in Memphis, by Selected Subgroups, 2011-12 50% 40% 30% 20% 10% 0 3% 5% 14% 16% 1% 4% 30% 27% 17% 12% White Latino Black 40% 46% Elementary students without disabilities Elementary students with disabilities Secondary students without disabilities Secondary students with disabilities In 2011-12, Memphis suspended students at a rate that was about double the national average. The highest suspension rate in Memphis was for secondary-level Black males with disabilities (53%). However, Memphis also showed some progress. To fully understand the district s discipline rates, it should be noted that suspension rates declined for all students at both the elementary and secondary levels from 2009-10 to 2011-12. Moreover, because the decline was greater for Black students than for White students, the discipline gap between Blacks and Whites narrowed slightly. As the Memphis example demonstrates, every district has a unique story. In addition to the selected profiles, this report features districts that showed improvement to demonstrate that large-scale change is not just an academic pipe dream, but that it is happening in hundreds of districts across the nation. 8

This national overview and the district profiles are two places where we combine our elementary and secondary analyses. Otherwise, this report and the companion spreadsheets divide the analysis into two sub-reports, elementary and secondary. At each level we provide a more in-depth look at national disparities, distribution of suspension rates, and trends, followed up with a more detailed district-level analysis that highlights the largest, highest-suspending, and most improved districts. In the appendix, we also provide the first-ever national analysis of suspension in K-8 schools. But first we note some critical background facts and details on the construction of this report. The Consensus Is That Frequently Suspending Children From School For Minor Offenses Is Not Educationally Sound. 12 The book, Closing the School Discipline Gap, the most recent collection of research on this issue, lays out a strong basis for this assertion (Losen, 2015). The book also provides new empirical evidence that demonstrates that there are more effective ways than exclusion from school to ensure that schools provide safe and healthy learning environments. Ultimately, the new research shows that the means for closing the school discipline gap are not only available, but they are also already being employed successfully by school districts across the country. These findings are bolstered by the empirical district-level data in this report, which shows that lower-suspending districts outnumber the highsuspending districts at both the elementary and secondary levels, and for every racial/ethnic group. High Suspension Rates Do Not Improve Learning Conditions Many readers believe that schools with high suspension rates are boosting the achievement of the students who don t misbehave. However, research on the state of Indiana, which controlled for poverty and race, found that lower-suspending schools had higher achievement rates (Skiba, 2014). Similarly, a study that tracked every middle school student in Texas over six years and controlled for over 80 variables found no academic benefits in schools with higher suspension rates (Fabelo, 2011). And, most notably, a recent study of the Denver Public Schools, where a concerted effort was made to improve the school climate systemically by implementing restorative practices, found that, as suspension rates came down, the racial discipline gaps narrowed and test scores rose consistently at all grade levels in nearly every subject for six consecutive years (Gonzalez, 2015). Finally, a recent study that tracked the effect of high suspension rates on 17,000 individual students who were never themselves suspended found that, over three years, the high rates appeared to lower their math and reading scores (Morris & Perry, 2014). Together these findings dispel as myth the common assertion that you must kick out the bad students so the good students can learn. The availability of viable alternatives to frequent suspension makes the findings about the highsuspending schools and districts found in every state across the country even more disturbing. Put simply, high suspension rates and large disparities are not justifiable. Safe Schools Embrace Positive Discipline For readers who may object to discipline reform on the basis of safety concerns, we call attention to the safety benefits that result from reducing suspension rates. For example, a study on the Chicago Public Schools shows that schools serving students from the highest-crime neighborhoods had a wide range of safety ratings (Steinberg, 2015). In these schools, higher safety ratings were predicted by the levels of teacher-student engagement and teacher-parent engagement. Some of the high-scoring schools serving students from the highest-crime areas felt as safe to both teachers and students as many serving students from the lowest-crime neighborhoods. Equally important, after controlling for demographics, the Chicago 9

schools that felt safer also used exclusionary discipline much less frequently than the schools that ranked low on safety (Steinberg, 2015). Connecting the dots, one can imagine how frequently relying on suspension for minor offenses could harm the positive relationships between teachers and students that foster a safe environment. Moreover, having large numbers of unsupervised youth hanging out in high-crime neighborhoods during school hours can increase the likelihood that students who are suspended will turn to delinquent activity and join gangs. As mentioned at the outset, extensive research confirms that excessive exclusionary discipline contributes to the risk for dropping out and for juvenile delinquency (Balfanz, 2015; Schollenberger, 2015). This likely makes schools, and the communities in which they are located, less safe (Finn, 2015). Most recently, the Council of State Governments Juvenile Justice Center facilitated the development of consensus among law enforcement officials, judges, administrators, teachers, researchers, and community groups. After convening to discuss school discipline issues for more than two years, the group recommended reducing reliance on suspension as part of broader efforts to improve the school climate and community safety (School Discipline Consensus Report, 2014). When discussing safety, it is also important to note that the data in this report do not include expulsions. Expulsions may also be excessive, but they are typically used in response to weapons and drug possession, and to address the most serious and violent behavior. In the Civil Rights Data Collection (CRDC), the source of all the data in this report, the nearly 3.5 million students suspended outnumber those expelled (130,000) by about 27 to 1. 13 Although the CRDC counts the number of students suspended but does not count the number of suspensions or provide information on the reasons for suspension, analysis of state data where these numbers are collected (Texas, Connecticut, Massachusetts, Ohio, North Carolina, and others) consistently shows that students are suspended most often for minor nonviolent violations of school codes of conduct, not unlawful or dangerous behavior. 14 In our own recent report on the reasons for suspensions in California, the data show that the most frequent reasons for removal were a range of minor misbehaviors that the state code labels Disruption and Willful Defiance (Losen & Martinez, 2014). This introduction emphasizes that concerns about excessive use of suspension are based on evidence that it is harmful to all students in high-suspending schools. In 2011-12 alone, approximately 1.1 million White students were suspended out of school compared to approximately 700,000 Latino and 1.2 million Black students (CRDC, 2014). Thus, reducing the harm done to students by limiting suspension to a measure of last resort will benefit millions. There Is Growing Concern about the Profound Disparities in Suspension Rates The main body of this report documents gross disparities in the use of out-of-school suspension experienced by students with disabilities and those from historically disadvantaged racial, ethnic, and gender subgroups. The egregious disparities revealed in the pages that follow transform concerns about educational policy that allows frequent disciplinary removal into a profound matter of civil rights and social justice. This implicates the potentially unlawful denial of educational opportunity and resultant disparate impact on students in numerous districts across the country. On the other hand, it is worth noting that there has been increased attention to and action on this issue across the nation since 2011 (Losen, 2015). Thus we urge readers to treat the data as snapshots of the very recent past. For example, if a district has changed its policies and practices in the last two years, those changes are not reflected in these data. Moreover, others may be pursuing changes for the first time this year. 10

In January 2014, the U.S. Justice and Education departments issued joint guidance on racial disparities in school discipline to the leaders of our nation s public schools. The clear message presented in the guidance is that school administrators must examine their data and discipline policies and practices, and undertake efforts to close the discipline gap where unjustifiable disparities are found (DOJ & DOE, 2014). While this legal and moral obligation to eliminate racial disparities is not new, this guidance is the first joint federal effort to explicitly call upon school leaders to take immediate action. 11

Methods Important notes on the data and methods. The CRDC is the only source providing data from every large district in the nation that also goes as far back as the 1970s. To get a more accurate sense of the use of exclusionary discipline, and to track changes over time, this report focuses on out-of-school suspensions. Future reports will look at corporal punishment, expulsion, school-based arrests, and referral to law enforcement. We will also cover in-school suspensions, for which data were first collected in 2009-10, in future reports. While the raw data we used for this report are available to any researcher and can be replicated, the analysis broken down by elementary and secondary levels is currently found only here. The source of the data and calculation method. This report uses the CRDC s data on students receiving either just one out-of-school suspension or more than one, and combines them to look at the unduplicated count of students receiving one or more out-of-school suspensions. For example, the Office of Civil Rights reports X students suspended once and Y students suspended two or more times. We add these two numbers and then divide the sum (suspended one or more times) by the total student enrollment. This basic calculation produces the percentage of students suspended. Throughout this report, we refer to this as the percent suspended, the suspension rate, or the risk for suspension. These rates are based on the unduplicated number of students suspended, and should not be confused with the number of suspensions, which are distinct counts of disciplinary actions and are not collected by the CRDC. In the appendix, readers will find a complete description of the calculations, data-cleaning methods, and data issues and limitations. Like all suspension data in the public domain, all the reported values in this report are rounded up or down to protect against disclosing personally identifiable student information. Because, as researchers, we were granted access to the actual counts, prior to publishing all enrollment counts and suspension counts were always rounded to the nearest 5. These rounded numbers are available in the companion spreadsheets. Such rounding typically has little impact on the suspension rates we calculated, but when enrollment numbers are low, rounding can distort the findings. Therefore, we have identified all instances where rounding has had a serious impact. In the text of this report, any distortions from the actual unrounded values are within two percentage points, plus or minus, and we only included trends whose direction remained true after rounding. 15 Further, to conform to the IES requirements for using the unrounded data in our analysis, all published counts of the number of schools and districts were rounded to the nearest 5. To avoid distortion due to rounding, we only report federal and state level counts of the numbers of high and lower-suspending schools. 12

This report provides the most accurate trend analysis of suspension rates at the elementary and secondary levels for this period. We describe the trends for each subgroup, as well as changes in the Black/White and Latino/White discipline gaps over time. We did not track the district trends for Asian Americans or Hawaiians/ Pacific Islanders because of issues with data quality and inconsistent definitions of these two groups. Rates for American Indians often are not included in the featured analysis because of concerns about data quality. However, the data for these three groups are highlighted in several places in this report, and all are presented in detail for every district in the companion spreadsheets. In some cases, we found conflicting data on district or state websites. Some large districts, such as New York City were removed from the analysis but can be found in the error tab in our spreadsheets. Full explanations are provided in the appendix. 13

Elementary School Suspension Rates Nationally: As the average national suspension rate for students in elementary school (2.6%) shows, the small number of students suspended out of school at least once is substantially lower than the K-12 suspension rate, because younger children are subjected to exclusionary discipline less frequently. That said, these rates have increased significantly since the early 1970s, following the general pattern of suspension rates. The increased use of suspension has come with an increase in the racial gap, with more students of color being suspended at the elementary level as well. The experiences of students in schools and districts vary tremendously from one district to the next, and often within the same school district. Although the suspension rate and racial gap in 2011-12 remained similar to what we observed in 2009-10, many schools and districts have substantially lowered their reliance on out-of-school suspensions, whereas others have increased their use. This variation raises questions about whether the high and disparate rates are necessary, or justifiable, at the elementary level. Elementary and preschool trends: The district trends we observe thus far at the elementary level may reflect a conscious effort by some district leaders to curtail out-of-school suspensions for elementary school students. However, the 2011-12 data do not reflect some recent and notable developments in elementary school suspension. Most notable is the new California legislation requiring the statewide elimination of some of the catch-all minor offense categories, known as Disruption or Willful Defiance, as grounds for suspension in grades K-3 (Cal. Education Code. 48900, 2014). Los Angeles eliminated this punishment category in 2012-13 for all students at every grade level. While recent developments suggest that elementary suspension rates will decrease, the U.S. Department of Justice shined a light for the first time on what many believe is the growing use of suspension for preschool children. We also find a great deal of disproportionality by race/ethnicity and gender in this category (Data Snapshot, 2014). However, it is difficult to ascertain whether suspension of preschoolers is on the rise or decline, because 2011-12 is the first time the data were collected at this level. Furthermore, the number of public preschools that reported data, preschool enrollment, and the number of suspended preschoolers is much lower than the number of elementary schools (K-5 or K-6) that did so. For these reasons, no analysis of the preschool data by district was included in this report. Racial gaps remain the same: Nationally, most subgroups at the elementary school level saw an increase in their risk for out-of-school suspension since 2009-10, but it was less than one percentage point. Given the likely harm caused by suspensions, it is no real indication of progress if the racial gap between Blacks and lower-suspended subgroups narrowed only because these subgroups had a larger increase in suspension rates than Black students. Changes in the racial gap at this level are not featured here, but they can be found in the trend companion spreadsheet for every district that reported data in 2009-10 and 2011-12. However, more substantial changes were observed in many districts and secondary schools where suspension rates are much higher. 14

Elementary school suspension rates have a broad scope: Many schools and entire districts did not suspend any elementary students out of school in the 2011-12 school year. Table 1 below provides a sense of the suspension rate distribution for elementary students by district for each racial group that had at least ten elementary students in a district. The graph and table below display the scope of district suspension rates for each racial/ethnic group at the elementary level. Figure 6. The Distribution of Elementary Suspension Rates, 2011-12 National Distribution of All Districts By Suspension Rate For Each Racial/Ethnic Group 100% 90% Over 10% 80% 7.5 to 10% 70% 5 to 7.5% 60% 50% 2.5 to 5% 40% 0 to 2.5% 30% 20% 10% 0 Black Latino White American Indian Asian American HI/Pac. Isl. Table 1. The Distribution of Elementary Suspension Rates, 2011-12 Elementary Level: Distribution of District Suspension Rates and Percent of All Districts by Race/Ethnicity Black Latino White Am. Indian Asian HI/Pac. Islander All 2.5 >2.5 and 5 >5 and 7.5 >7.5 and 10 >10 Total Districts* 58% 11% 9% 9% 3% 8% 14% 5,825 84% 5% 3% 3% 4% 2% 1% 86% 80% 98% 93% 75% 13% 4% 1% 0% 15% 5% 3% 1% 1% 2% 0% 2% Percentages rounded to whole numbers. All rows will not add up to 100% due to rounding. 1% 10,215 0% 4,530 1% 16 8,055 3% 2,520 915 10,595 *District numbers rounded to nearest 5. In most of the nation s school districts, elementary suspension rates are less than or equal to 2.5% of the student enrollment for each racial and ethnic group. The data suggest that the excessive and disparate use of suspension is occurring at an egregious rate in elementary schools in a relatively small subset of school districts. This review supports the argument that the excessive use of discipline violates the norm for elementary students. We see a similarly wide distribution at the state level. 15

Table 2. States Elementary Rates, Ranked by the Black/White Gap, 2011-12 State All Students 2011-12 American Indians Asian Americans Hawaiian Pacific Islanders Black Students Latino Students White Students Black/ White Gap in Percentage Points MO 3.8% 0.6% 0.5% 0.0% 14.3% 1.9% 1.8% 12.5 WI 2.0% 2.7% 0.4% 0.0% 12.2% 2.0% 1.0% 11.2 IN 3.5% 1.6% 0.3% 1.8% 12.8% 2.7% 2.1% 10.6 MI 3.6% 3.1% 0.4% 1.0% 12.1% 2.6% 2.1% 10.0 OH 2.9% 0.7% 0.2% 0.0% 11.1% 2.0% 1.7% 9.5 DE 4.9% 0.0% 0.0% * 10.6% 2.9% 2.2% 8.4 PA 2.0% 0.5% 0.2% 0.0% 9.0% 3.4% 1.1% 8.0 AR 3.8% 0.5% 0.6% 0.4% 10.2% 1.3% 2.3% 7.9 NE 1.9% 4.9% 0.6% 0.0% 8.9% 1.4% 1.1% 7.8 TN 3.1% 1.8% 0.3% 0.0% 8.6% 1.0% 1.3% 7.3 OK 3.0% 2.1% 0.5% 0.0% 9.2% 2.7% 2.2% 7.0 RI 3.1% 5.4% 1.8% * 8.4% 5.2% 1.5% 6.9 FL 5.1% 3.6% 0.6% 2.5% 9.8% 4.3% 3.3% 6.5 IL 1.8% 0.5% 0.1% 0.0% 7.1% 1.1% 1.0% 6.1 IA 1.2% 0.5% 0.2% 0.0% 6.7% 1.2% 0.8% 6.0 NC 3.5% 6.0% 0.5% 1.2% 7.8% 1.9% 1.9% 5.9 MN 1.4% 4.4% 0.4% 0.0% 6.4% 1.4% 0.7% 5.7 CA 2.6% 3.9% 0.9% 2.0% 7.9% 2.4% 2.3% 5.6 TX 2.1% 0.8% 0.3% 0.3% 6.6% 1.7% 1.2% 5.4 KS 1.6% 1.2% 0.5% 1.3% 6.5% 1.6% 1.0% 5.4 SC 4.1% 2.0% 0.5% 0.0% 7.6% 1.8% 2.2% 5.3 VA 2.6% 1.7% 0.2% 1.6% 6.7% 1.3% 1.5% 5.3 GA 3.3% 0.9% 0.5% 1.8% 6.4% 1.5% 1.6% 4.8 LA 4.3% 5.0% 0.5% 0.0% 6.9% 1.4% 2.2% 4.7 WA 2.4% 3.7% 0.6% 2.2% 6.8% 2.6% 2.1% 4.6 CO 2.0% 3.1% 0.6% 0.0% 6.1% 2.4% 1.5% 4.6 CT 1.3% 0.0% 0.0% 0.0% 5.0% 2.5% 0.4% 4.6 MS 4.8% 1.0% 0.6% * 7.1% 1.6% 2.5% 4.5 AL 3.2% 1.1% 0.1% 4.4% 6.2% 1.0% 1.7% 4.5 AZ 2.4% 3.8% 0.6% 2.1% 6.5% 2.1% 2.1% 4.4 NJ 1.2% 0.0% 0.1% 0.0% 4.7% 1.3% 0.5% 4.2 OR 2.3% 3.3% 0.6% 1.2% 6.2% 1.9% 2.3% 3.9 WV 2.9% 0.0% 0.0% * 6.5% 1.4% 2.7% 3.8 MA 1.4% 1.1% 0.2% 0.0% 4.1% 3.2% 0.7% 3.4 ME 1.1% 0.0% 0.0% * 3.9% 1.1% 1.0% 2.9 KY 1.4% 1.3% 0.2% 0.0% 3.9% 0.7% 1.0% 2.9 UT 1.0% 2.4% 0.7% 0.7% 3.6% 1.6% 0.9% 2.8 AK 2.1% 3.0% 0.7% 2.3% 4.3% 1.9% 1.8% 2.6 WY 1.6% 2.8% 0.0% 0.0% 3.8% 1.6% 1.4% 2.4 NV 1.6% 2.9% 0.5% 1.2% 3.9% 1.3% 1.6% 2.4 16

State All Students 2011-12 American Indians Asian Americans Hawaiian Pacific Islanders Black Students Latino Students White Students Black/ White Gap in Percentage Points MD 1.7% 0.4% 0.2% 0.4% 3.1% 0.8% 1.0% 2.1 SD 1.1% 2.8% 0.0% * 2.8% 0.9% 0.8% 2.0 NM 1.9% 1.9% 0.3% 0.0% 3.2% 2.1% 1.4% 1.7 VT 1.5% 0.0% 0.9% * 2.8% 3.8% 1.4% 1.4 MT 2.2% 6.8% 0.0% 0.0% 2.0% 0.8% 1.4% 0.5 NH 1.1% 0.0% 0.2% * 1.5% 1.9% 1.1% 0.5 ID 1.2% 4.0% 0.0% 0.0% 1.4% 1.2% 1.2% 0.2 ND 0.5% 2.1% 0.0% 0.0% 0.3% 0.0% 0.3% 0.0 Highest-suspending states for each subgroup are bolded. Asterisk means that there were less than 100 students reported as enrolled in the state. Hawaii and New York were removed because of data reporting errors. Are certain states suspending elementary school students at significantly higher rates? Before 2011-12, several states and districts had implemented initiatives to reduce suspension rates, and some focused on elementary schools in particular. In Maryland, for example, the state passed a law in 2004 that requires any elementary school that suspends over 10% of its total enrollment to engage in corrective action. 17 Connecticut passed a law in 2009, which was implemented in 2011, aimed at eliminating outof-school suspensions except as a measure of last resort. 18 Both states are on the lower end of the overall state rankings on suspension. Maryland also had 305 lower-suspending elementary schools statewide and just 80 high-suspending schools, for a ratio of about 3.8 to 1, which is much better than the national ratio of 2.2 to 1. The number of high and lower suspending schools for each state can be found in the companion spreadsheet. On the other hand, Florida had the highest overall suspension rates for elementary students. Furthermore, the number of high-suspending elementary schools outnumbered the lower-suspending elementary schools in only seven states Delaware, Florida, South Carolina, Louisiana, North Carolina, Georgia, and Mississippi. In Missouri the numbers were just slightly higher for lower-suspending schools. At 12.5 more Black elementary students than White elementary students suspended per every 100 enrolled, Missouri s Black-White discipline gap was the widest in the nation at the elementary level, and it also had the highest Black elementary suspension rate of any state. It is worth noting that Missouri schools had White elementary suspension rates on a par with the national average for all students. Moreover, Black elementary students in Missouri are suspended at higher rates than the state s White secondary school students. Elementary School Suspension at the District Level Suspension in the largest districts: In many of the nation s largest districts, the elementary school risk for out-of-school suspension is similar to the national average (2.6%). The table below shows the risk for suspension for all elementary school students in ten of the nation s largest districts for which we had reliable data in 2009-10 and 2011-12. 19 While an equal number of the selected large districts showed declines as showed increases in 2011-12, they all had suspension rates below 4% of total enrollment, and half had a rate below 2%. 17

Table 3. Elementary Suspension Rates in Ten of America s Largest School Districts District State 2009-10 2011-12 Trend in Percentage Points LOS ANGELES UNIFIED CA 1.5% 0.8% -0.7 CLARK COUNTY SCHOOL DISTRICT NV 1.0% 1.5% +0.5 HOUSTON ISD TX 4.3% 3.8% -0.5 FAIRFAX CO PBLC SCHS VA 0.8% 0.3% -0.5 DALLAS INDEPENDENT SCHOOL DISTRICT TX 6.0% 3.2% -2.8 GWINNETT COUNTY PUBLIC SCHOOLS GA 1.7% 2.1% +0.4 WAKE COUNTY SCHOOLS NC 2.0% 1.9% -0.1 MONTGOMERY COUNTY PUBLIC SCHOOLS MD 0.3% 0.6% +0.3 CHARLOTTE-MECKLENBURG SCHOOLS NC 0.6% 3.3% +2.7 PRINCE GEORGE S COUNTY MD 2.5% 2.8% +0.2 Although not included in Table 3, the data in the companion spreadsheet show that the risk for elementary-level suspension for Black students was below 10% in all of the above large districts. Although most of these districts would not meet our criteria for lower-suspending for every racial group, it is noteworthy that none met the criteria for high-suspending either. Next we feature some of the nation s highest-suspending districts. The Highest-Suspending Districts Have Alarming Rates of Suspension Each of the districts below had at least 1,000 elementary school students and each suspended over 20% of its total elementary enrollment at least once in 2011-12. Table 4. 10 Highest-Suspending Districts for ALL Elementary School Students District State All OSS 2009-10 All OSS 2011-12 Trend in Percentage Points PONTIAC CITY SCHOOL DISTRICT MI 39.6% 31.7% -7.9 ST. LOUIS CITY MO 17.5% 29.1% +11.6 TROTWOOD-MADISON CITY OH 22.1% 25.4% +3.3 WOODLAND HILLS SD PA 19.3% 23.8% +4.5 NORMANDY MO 18.1% 21.7% +3.5 RIVERVIEW GARDENS MO 22.9% 21.4% -1.5 EAST CLEVELAND CITY SCHOOL DISTRICT OH 22.9% 21.2% -1.7 EAST DETROIT PUBLIC SCHOOLS MI 10.0% 20.9% +10.9 YORK CITY SD PA n/a 20.6% n/a TAYLOR FL n/a 20.5% n/a Three of the highest-suspending districts for elementary school students were in Missouri, and two each were in Ohio, Michigan, and Pennsylvania. Given the unrest in Ferguson, Missouri, it is worth noting that Michael Brown reportedly attended high school in Normandy, Missouri, which, like Riverview Gardens (cited above as one of the highest-suspending districts in the nation for elementary students), is part of the greater St. Louis metropolitan area. 20 St Louis is also on the list. 18

In five of the ten districts with the highest elementary suspension rates, rates have risen significantly since 2009-10: For example, rates rose by over nine percentage points in East Detroit, Michigan and St. Louis. However, rates decreased significantly (by eight points) in the highest-suspending district, Pontiac, Michigan, and by over 1 point in Riverview Gardens, Missouri, and East Cleveland, Ohio. Students with disabilities were also suspended at a high rate in these districts, between 29% and 47% of their enrollment, and at a substantially higher rate than their peers without disabilities. Black male suspension rates in elementary school in these highest-suspending districts ranged from 31% to 56%. Demographics are not destiny: Demographically, the seven highest-suspending districts all had majority Black enrollment, although the range was from 26% to 99% Black. Only one district, Taylor, Florida, was majority White, at 67%. In searching for districts to compare with those above, we found 51 districts from across the nation with at least 25% Black enrollment and at least 1,000 elementary school students, where elementary suspension rates were under 3% for all students and under 3% for Black students (range 2.78% to 0.22%). These included large urban districts, such as Broward County, Florida, and Boston, Massachusetts. There would have been many more districts in this category, but we eliminated all that reported suspending no elementary students or no Black elementary students. Although it is not uncommon for elementary schools to have zero suspensions, we suspect that some such districts failed to report their data. Black enrollment in these low-suspending comparison districts ranged from 25% to 96%. Furthermore, we found that suspension rates for elementary students in 35 of the lower-suspending districts had declined since 2009-2010. In other words, readers would be wrong to assume that something about the behavior of Black elementary students requires greater use of suspension. To the contrary, these data, along with several studies that tracked behavior ratings of students as well as disciplinary outcomes (Skiba, 2015; Finn, 2015; Fabelo, 2011) suggest that Black students are punished more harshly and more often for subjective minor offenses. Instead, researchers conclude that school policies and practices more than differences in behaviors, predict higher suspension rates (Skiba, 2015; Finn, 2015; Fabelo, 2011; Bradshaw, 2011). The large number of high-suspending elementary districts in Missouri prompted further investigation into these patterns. Did Missouri s districts look different from other states? Like all states, within the state, Missouri s suspension rates and racial gaps vary dramatically from one district to the next. When we looked at the 51 districts in Missouri that enrolled at least 1,000 elementary school students and at least 100 Black students, we found that some of these districts relied on suspension far more than others. Four districts St. Louis, Kansas City, Normandy, and Riverview Gardens had rates above 20% for Black elementary students. These four districts also enrolled about one-third of all the state s Black elementary school students. Although the racial gaps were large 28 more Black than White students suspended per 100 enrolled in St. Louis the four districts were also among the highest-suspending for White elementary school students. Another 11 districts had Black elementary suspension rates above 10% but below 20%. It deserves mention that about 4/5 ths of the state s Black enrollment attends elementary school in a district with suspension rates above 10% for Black elementary students. However, 15 districts had rates between 5% and 10% for Black elementary students, and 21 districts serving 8,685 Black elementary students had suspension rates below 5% for those Black students. In other words, 19

even in Missouri, far more school districts had Black elementary suspension rates that were close to the national average for all elementary students (2.6%) than had high suspension rates (above 10%). Finally, although a full-scale review of each state s code of conduct is beyond the scope of this report, we found nothing in Missouri s code to suggest there was a particularly harsh state policy. Elementary School Students with Disabilities Are at Profound Risk in Some School Districts When we looked at districts with at least 100 elementary students with disabilities and at least 1,000 students enrolled, we found shocking suspension rates for students with disabilities in 37 school districts more than 25%! We list the top ten in the table below and compare the rates for students without disabilities (SWOD) to those with disabilities (SWD) to give a clear picture of the broad scope of the problem. Table 5. Students with Disabilities Compared to Those without in Highest-Suspending Elementary School Districts District State OSS for Elementary SWOD OSS for Elementary SWD The Percentage Point Difference CLOVIS UNIFIED CA 3.8% 56.9% 53.1 TROY CITY AL 0.5% 40.9% 40.4 CLOVER PARK SCHOOL DISTRICT WA 9.6% 46.8% 37.3 MADISON CO SCHOOL DIST MS 1.5% 38.1% 36.6 NAVASOTA ISD TX 5.6% 39.1% 33.5 EDGEFIELD 01 SC 0.6% 27.5% 26.9 AVOYELLES PARISH LA 7.2% 33.9% 26.7 LAKE FL 6.7% 28.7% 22.0 FREDERICKSBURG CITY PBLC SCHS VA 6.6% 28.6% 22.0 HANOVER AREA SD PA 7.8% 29.7% 21.9 Disability rights are in jeopardy: The huge differences observed between students with and without disabilities are shocking. In a separate analysis we found that students with an emotional disturbance or with significant learning disabilities had the highest risk for suspension among students with disabilities (Losen, Ee, Hodson, & Martinez, 2015). The inference from that study is worth repeating: One plausible explanation is that students with such disabilities are more likely to misbehave because of their disability. However, suspending students for behavior that is a manifestation of their disability is unlawful. Moreover, schools are obligated to determine if the disability is causing the misbehavior... It is their legal responsibility not to suspend children because of their disabilities. (Losen et al., 2015, p. 99) 20